Magnetic resonance imaging (MRI) of the bowel is an increasingly used modality to evaluate patients with Crohn’s disease. The Montreal classification of the disease behavior is considered as an excellent prognostic and therapeutic parameter for these patients. In our study, we correlated the behavior assessment performed by a radiologist based on MRI with the surgeons' clinical assessment based on the assessment during abdominal surgery.
We evaluated 76 patients with Crohn’s disease, who underwent bowel resection and had an MRI within 4 weeks before surgery. Radiological behavior assessment was performed by 2 radiologists based on MRI. Behavior was classified into B1 (nonstricturing and nonpenetrating), B2, and B3 (penetrating) disease. Surgical assessment was done by the same surgeon, who performed all bowel resections, based on intraoperative findings and histologic results.
The surgical assessment identified 4 patients (5%) as B1, 16 patients (21%) as B2, and 56 patients (74%) as B3. In 97% (n = 74) of all patients, the intraoperative and radiological assessment were identical with interobserver agreement of 0.937. In one case, B2 was radiological considered as B1, and in another case, B3 was diagnosed as B2. The diagnosis of a stricture had the highest sensitivity of 96%, whereas the detection of inflammatory mass showed the lowest sensitivity of 81%. Abscesses had the lowest positive predictive value of 68% with a specificity of 88%. Best correlation was found for fistulae (0.895).
MRI represents an excellent imaging modality to correctly assess the Montreal classification–based disease behavior in patients scheduled for bowel resection with Crohn’s disease.
Article first published online 7 March 2013
*Department of Radiology, University Medical Center Regensburg, Germany
†Department of Surgery, Marienhospital, Gelsenkirchen, Germany
‡Department of Internal Medicine I
§Department of Surgery, University Medical Center Regensburg, Germany.
Reprints: Andreas G. Schreyer, MD, MBA, Department of Radiology, University Medical Center Regensburg, Franz-Josef-Strauss-Allee 11, 93051 Regensburg, Germany (e-mail: email@example.com).
The authors have no conflicts of interest to disclose.
Received July 11, 2012
Accepted July 17, 2012